Technical Abstract: The Water Erosion Prediction Project (WEPP) model simulates furrow irrigation hydrology, equates furrow erosion with rill erosion processes, and uses soil parameters derived from rainfall simulation and averaged climate data. Our objective was to evaluate the WEPP model for furrow irrigation by comparing predicted infiltration, runoff and soil erosion with field measurements from two separate studies. One study tested tillage treatment effects on an 85-m long field segment, which was approximately the upper third of the field. The other study tested inflow rate effects on 204- and 256-m long fields. Predicted annual runoff was within 5% and infiltration was within 25% of measured values for the 85-m long field segment. On the two longer fields with four overland flow elements, predicted annual infiltration was 50 - 80% greater than measured for one field and 15% less than measured for the other field. Consequently, predicted annual runoff was 1/3 to 2 times the measured values. No soil loss was predicted when using WEPP defined critical shear (3.5 Pa) and rill erodibility (0.0215 s m-1) parameters for Portneuf silt loam (coarse-silty, mixed, mesic Durixerollic calciorthid). WEPP over-predicted annual soil loss by 2 - 15 times when critical shear and rill erodibility were adjusted to 0.72 Pa and 0.0021 s m-1 respectively. The model also failed to predict any erosion during irrigations with as high as 70 kg m-1 soil loss. Some of the erosion prediction error may result from WEPP using a linear relationship between detachment capacity and hydraulic shear. Power functions had 0.03 - 0.22 higher coefficients of determination than linear functions for 26 of the 33 WEPP soils.